We buy plans the same way you would, use them around Osaka and on trips across Kansai, and record the same measurements every time. Prices are checked against each provider’s own site, and every price on this site shows the date it was checked.
What we measure
| Measurement | How | Why it matters |
|---|---|---|
| Download and upload speed | Speedtest by Ookla, three runs per location, median kept | Shows whether maps, video calls and uploads will feel quick |
| Ping (latency) | Same Speedtest runs | High ping makes video calls and navigation lag. Some travel eSIMs route traffic through another country, which shows up here |
| Setup time | Minutes from purchase to installed, then from switching on in Japan to first working data | Tells you how much of your arrival day it takes |
| Coverage | Notes on dropouts underground, on trains and in the countryside | City speed tests alone hide where connections fail |
| Support | One real question sent to support, reply time recorded | You only find out how good support is when something breaks |
Where we test
Every plan goes through the same route, so results can be compared:
- City: street level in Umeda, Osaka, on a weekday morning.
- Underground: on the Midosuji subway line between Umeda and Namba.
- Train: on the Shinkansen between Shin-Osaka and Kyoto.
- Countryside: Koyasan in Wakayama, as a stand-in for temples, mountains and rural trips.
How results appear on reviews
Each review has a “How it performed in Osaka” section. Until a plan has been through the route, that section says so plainly, and the review doesn’t claim any first-hand testing. When results exist, they appear exactly as in the example below, with the date of the test and the phone used.
What we don’t do
- We don’t let providers see or change a review before it’s published.
- If a provider gives us a plan for free, the test says so next to the results.
- An affiliate commission never decides a ranking. Providers we earn nothing from are compared the same way. How we make money
How often we re-check
Prices and plan details are re-checked against provider sites every quarter, and each one shows its “checked” date. Speed tests run on the same quarterly cycle, and sooner if a provider changes the network it uses or readers tell us performance has changed.
Example: what a finished test looks like
These two tests are invented, with made-up provider names. They show the layout and the kind of detail each real test will include.
Example data with made-up providers, to show what a finished test looks like. These are not real results.
Example eSIM A 10GB / 30 days
docomo- Median download
- 37.5 Mbps
- Median upload
- 7.5 Mbps
- Median ping
- 48 ms
- Setup
- 4 min to install, 1 min to first data
| Where | Download | Up | Ping |
|---|---|---|---|
| Umeda, Osaka (street level) City, 08:40 | 142 Mbps | 21 | 38 ms |
| Midosuji subway line, Umeda to Namba Underground, 09:10 | 48 Mbps | 9 | 44 ms |
| Shinkansen, Shin-Osaka to Kyoto Train, 12:15 | 27 Mbps | 6 | 52 ms |
| Okunoin, Koyasan (Wakayama) Rural, 15:30 | 9 Mbps | 2 | 61 ms |
- QR code install. Data came on as soon as roaming was switched on.
- Signal held through the subway tunnels between stations.
- Dropped to no service for about 4 minutes on the Koyasan cable car.
Support: contacted by email, replied in 6 hours. Asked how to top up. Clear answer the same day.
Example eSIM B Unlimited / 7 days
SoftBank- Median download
- 18 Mbps
- Median upload
- 4 Mbps
- Median ping
- 145.5 ms
- Setup
- 11 min to install, 6 min to first data
| Where | Download | Up | Ping |
|---|---|---|---|
| Umeda, Osaka (street level) City, 08:35 | 63 Mbps | 11 | 128 ms |
| Midosuji subway line, Umeda to Namba Underground, 09:05 | 22 Mbps | 5 | 141 ms |
| Shinkansen, Shin-Osaka to Kyoto Train, 12:20 | 14 Mbps | 3 | 150 ms |
| Okunoin, Koyasan (Wakayama) Rural, 15:25 | 3 Mbps | 1 | 188 ms |
- App install. Needed to restart the phone before data worked.
- High ping everywhere: traffic appeared to route through an overseas server, which made video calls lag.
Recording a test
For each plan, note these on the day. They map straight onto the results above:
- Date, phone model and OS version, and whether you paid for the plan
- Minutes to install, and minutes from landing (or switching on) to first data
- At each of the four locations: time of day, then download, upload and ping as the median of three Speedtest runs
- Anything that went wrong: no signal, slow spots, needing a restart, support contact
- Optional: screenshots of the Speedtest results